Lightroom Search

Lightroom Catalog Search

Search a Lightroom CC catalog through a SQLite+FTS5 index built from Adobe's internal msgpack-based catalog.

How It Works

Lightroom CC stores its catalog as a SQLite database containing msgpack-encoded document revisions. The indexer reads these revisions, extracts photo metadata (EXIF, keywords, ratings, etc.), and writes them into a searchable SQLite database with full-text search.

Two databases are involved:

1. Lightroom's catalog (read-only source) — the .mcat file inside the .lrlibrary bundle
2. The search index (created by the indexer) — a regular SQLite DB you query against

Catalog Location

The Lightroom CC catalog lives at:

~/Pictures/Lightroom Library.lrlibrary/<catalog-id>/Managed Catalog.mcat

The <catalog-id> is a UUID specific to each user's library (e.g., b67ffe07bd574f4b89f1f6194ebb8d3b). To find it:

ls ~/Pictures/Lightroom\ Library.lrlibrary/

There is typically one directory — that's the catalog ID.

Prerequisites

pip install msgpack

Indexing the Catalog

Use lightroom-index.py to build the search index. The script reads the Lightroom catalog and creates a SQLite+FTS5 database.

Configuration

Edit these constants at the top of lightroom-index.py to match the user's setup:

# Path to Lightroom's internal catalog database
CATALOG_PATH = os.path.expanduser(
    "~/Pictures/Lightroom Library.lrlibrary/<catalog-id>/Managed Catalog.mcat"
)

# Where original photo files live (optional, for --with-path resolution)
ORIGINALS_BASE = "/path/to/Lightroom/originals/<catalog-id>/originals"

# Where to write the search index
DB_PATH = "/path/to/lightroom.db"

Running the Indexer

# Full index (first time, or to rebuild)
python3 lightroom-index.py ingest --full

# Incremental update (only new/changed revisions since last run)
python3 lightroom-index.py ingest

The indexer tracks its last-processed revision sequence, so incremental updates are fast.

Searching

CLI Search

# Full-text search (searches filenames, keywords, camera info)
python3 lightroom-index.py search "sunset"

# With filters
python3 lightroom-index.py search "portrait" --rating 4 --year 2024
python3 lightroom-index.py search "*" --camera "Sony" --lens "85mm"
python3 lightroom-index.py search "*" --edited --rating 3
python3 lightroom-index.py search "*" --date 2024-06-15
python3 lightroom-index.py search "*" --date-from 2024-01-01 --date-to 2024-03-31

# Include file paths in results
python3 lightroom-index.py search "landscape" --with-path --limit 50

# Use "*" as query to match all (filter-only mode)
python3 lightroom-index.py search "*" --rating 5

Search Filters

Other Commands

# Library statistics (total assets, rating distribution, top cameras, years)
python3 lightroom-index.py stats

# Top keywords in the library
python3 lightroom-index.py keywords --top 30

# Random rated photo (optionally filtered)
python3 lightroom-index.py random --rating 4 --keyword landscape

Querying the Index Directly

For more complex queries, open the index database directly with SQLite:

import sqlite3
conn = sqlite3.connect("/path/to/lightroom.db")

Common Queries

-- Full-text search via FTS5
SELECT a.* FROM assets a
WHERE a.id IN (SELECT asset_id FROM assets_fts WHERE assets_fts MATCH 'sunset');

-- Top-rated photos from a specific year
SELECT filename, capture_date, rating, camera_model, lens
FROM assets WHERE rating >= 4 AND capture_year = 2024
ORDER BY rating DESC, capture_date DESC;

-- Photos by camera
SELECT filename, capture_date, rating
FROM assets WHERE camera_model LIKE '%A7R%'
ORDER BY capture_date DESC;

-- All keywords for a photo
SELECT k.keyword FROM keywords k WHERE k.asset_id = 'some-asset-id';

-- Photos with a specific keyword
SELECT a.filename, a.capture_date, a.rating
FROM assets a
JOIN keywords k ON a.id = k.asset_id
WHERE k.keyword = 'landscape';

-- Rating distribution
SELECT rating, COUNT(*) FROM assets GROUP BY rating ORDER BY rating;

-- Photos per year
SELECT capture_year, COUNT(*) FROM assets
WHERE capture_year IS NOT NULL GROUP BY capture_year ORDER BY capture_year;

-- Top cameras used
SELECT camera_model, COUNT(*) c FROM assets
WHERE camera_model IS NOT NULL GROUP BY camera_model ORDER BY c DESC LIMIT 10;

-- Edited vs original count
SELECT is_edited, COUNT(*) FROM assets GROUP BY is_edited;

Database Schema

See references/schema.md for the full schema. Key tables:

Key Asset Fields

Reference Files

• references/schema.md — Full database schema with all columns and indexes
• references/indexer.md — The complete indexer script for reference